e/o, aspect, how old?

Wrt the idea that preverb-induced aspect might go back to PIE times,
perhaps it is relevant that a few Russian perfective/imperfective
pairs show consistent e/o distribution:
vestí/vodít', veztí/vozít', nestí/nosít'
If so, the e-grade verbs of those pairs were 'born' perfective, and
the o-grade ones the result of de-preverbing.
On the other hand, against a general solution by an phonologically
driven é-o- rule militates the fact that by far not all Russian
preverbs are e-grade. Perhaps one should consider a solution à la
Germanic, with some composite verbs divisible (trennbar) and some not?


Torsten